From Zero to Market: Building a Resilient Web Solution in Just 7 Months

Business Challenges

Multi-Channel Settings Management

• Challenge:

  Managing settings across multiple domains individually.

• Solution:

  Developed a custom admin panel for centralized control of images, themes, and SEO/meta data per channel.

Advanced Product Variations & Reviews in Akeneo

• Challenge:

  Handling complex variations (color, size, material) and integrating customer reviews.

• Solution:

  Extended Akeneo with custom associations for variations and a dedicated review module linked to product data.

Content Page Management

• Challenge:

  Need for unified content editing across domains using global variables.

• Solution:

  Built a content management module in the admin panel for easy edits and dynamic content personalization.

Dynamic Data & Complex Queries

• Challenge:

  Managing large volumes of backend-driven UI data without performance issues.

• Solution:

  Designed optimized query structures and synced endpoints for real-time updates with minimal latency.

Scalability via Kubernetes

• Challenge:

  The existing infrastructure lacked dynamic scalability and efficient container management.

• Solution:

  Migrated to Kubernetes for automated deployment, scaling, and high application availability — a key part of our Kubernetes development and DevOps containerization strategy.

Large File Import Handling

• Challenge:

  Risk of failed imports due to large file size or system interruptions.

• Solution:

  Split imports into atomic tasks using Bull (Node.js) and Redis for retry logic and reporting.

Optimizing Search with ElasticSearch

• Challenge:

  PostgreSQL struggled with large datasets and flexible filtering.

• Solution:

  Integrated ElasticSearch for faster, scalable search, advanced filtering, and full-text query support.

Determine filterable attribute for products

• Challenge:

  PIM doesn't provide an out-of-the-box opportunity to determine which attributes should be used for a particular category, so that we could filtering products by them. Each product can have its own set of filterable attributes, and if not for any workarounds, we would have to query all the products and its filterable attributes (which would not be feasible task at all)

• Solution:

  We have introduced a flow where the admin should create a family (~set of attributes), which name corresponds to the title of the category, and specify the exact set of filterable attributes directly. So when we query a category, we can easily determine its own set of attributes without querying all the products.

Handle big import file import

• Challenge:

  We have a functionality when we have to handle big import files and create an database entity corresponding to each row. But the problem was that we could not guarantee that the import will finish , for example , before deployment or without other reason for breaking the connection.

• Solution:

  We separated the whole task to a set of atomic tasks (each of them were responsible to handle and create only one entity), used task queue (a Nodejs library Bull) with Redis as a broker. If one of more of the atomic tasks fails, we can set it to retry. When the last task is finished, we collect the report about the whole file and sent it to the user, knowing that the import has finished correctly without lost of data.